NBER WORKING PAPER SERIES HENRY AGARD WALLACE, THE IOWA CORN YIELD TESTS, AND THE ADOPTION OF HYBRID CORN Richard C. Sutch Working Paper 14141 http://www.nber.org/papers/w14141 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 June 2008 Thanks to Connie Chow and Hiroko Inoue for research assistance, to Susan B. Carter for critical advice, to Mason Gaffney for prodding questions that stimulated much further research, and to Norman Ellstrand for assistance with the plant biology. Financial support was provided by a National Science Foundation Grant: “Biocomplexity in the Environment, Dynamics of Coupled Natural and Human Systems.” Administrative support was provided by the Biotechnology Impacts Center and the Center for Economic and Social Policy at the University of California, Riverside. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer- reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications. © 2008 by Richard C. Sutch. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. Henry Agard Wallace, the Iowa Corn Yield Tests, and the Adoption of Hybrid Corn Richard C. Sutch NBER Working Paper No. 14141 June 2008 JEL No. N12 ABSTRACT This research report makes the following claims: 1] There was not an unambiguous economic advantage of hybrid corn over the open-pollinated varieties in 1936. 2] The early adoption of hybrid corn before 1937 can be better explained by a sustained propaganda campaign conducted by the U.S. Department of Agriculture at the direction of the Secretary of Agriculture, Henry Agard Wallace. The Department's campaign echoed that of the commercial seed companies. The most prominent hybrid seed company, Pioneer Hi-Bred Company, was founded by Wallace and he retained a financial interest while serving as Secretary. 3] The early adopters of hybrid seed were followed by later adopters as a consequence of the droughts of 1934 and especially 1936. The eventual improvement of yields as newer varieties were introduced explains the continuation and acceleration of the process. Richard C. Sutch Economics Department Sproul Hall University of California Riverside, CA 92521 and NBER [email protected] Every economist knows the story of hybrid corn. Zvi Griliches (1930-1999) made the adoption of hybrid varieties in the United States the exemplar for his model of technological diffusion [Griliches 1957b]. This classic article is taught in every graduate program and routinely receives over twenty Web of Science citations a year. As evidence of its iconic status, the article’s citation counts have been rising steadily over the last half century [Diamond 2004: 380]. Reproducing data originally collected by the Department of Agriculture’s Statistical Reporting Service on the percentage of corn acreage planted with hybrid seed, Griliches observed a lazy S-shaped cumulative logistic diffusion pattern. Figure 1 returns to the original data and extends the coverage to 1960 (the last date available). It illustrates the well-known shape. Hybrid corn (technically “double-cross inbred-hybrid corn”) was “invented” by Donald F. Jones in 1917-1918 and was developed and introduced on a trial basis in 1924 by Henry Agard Wallace. In the 1920s the Iowa Experiment Station began scientific field trials. Wallace’s hybrid was first entered in 1921. It won first place in 1924. It was first sold commercially in 1925. Competitors began sales in 1928. Widespread commercial adoption began in 1932 [Olmstead 2006: IV-9; Zuber and Robinson 1941: 589]. In 1933 about 0.1 percent of the nation’s corn acreage was planted to the new seed. By 1960, 96.3 percent of acreage was planted to hybrid varieties [USDA, Agricultural Statistics 1962, Table 46: 41; USDA, Track Records, April 2004: 19]. Griliches considered the adoption pattern displayed in Figure 1 to have been remarkably rapid [Griliches 1957b: 502]. Griliches’s explanation for the rapid and complete abandonment of open- pollinated corn in favor of the new hybrid varieties was based on a simple set of “stylized facts.” He considered hybrid corn superior to the traditional open-pollinated varieties from the beginning and suggested that that superiority was established in 1935 and persisted thereafter.1 The advantage of hybrids, according to Griliches, could be 1 The date 1935 is the year that acreage planted to hybrid corn exceeded ten percent of the total in the district at the heart of the hybrid revolution. Griliches chose ten percent “as an indicator that the Richard Sutch, Adoption of Hybrid Corn Page 2 of 26 objectively measured by the relative increase in yield over the open-pollinated corn [Griliches 1957b: 516-517]. He assumed that the new varieties required no significant increase in capital investment or annual inputs. According to this analysis the adoption process in a given district was one of disequilibrium transition [p. 503]. Griliches attributed the lags in the process to “imperfect knowledge.” It “takes time to realize that things have in fact changed” [p. 516]. The spread of hybrid corn geographically was slowed somewhat by the supply lags in developing and introducing hybrid varieties tailored to the specific soil type, weather conditions, and latitude of the peripheral regions.2 But even this process was rapid. Using the rule of thumb suggested by Griliches to mark the start of an adoption process as the date that ten percent of acreage was planted to hybrid corn, Iowa in 1936 was followed by Illinois, Indiana, and Wisconsin in 1937, by Minnesota in 1938, and by Ohio, Nebraska, and Missouri in 1939.3 See Figure 2. The development of hybrid corn and its rapid adoption were, nearly from the beginning, hailed as a triumph of twentieth-century biotechnology and one that carried with it enormous welfare benefits [Sprague 1946:101].4 In a chart that is perhaps even development had passed the experimental stage and that superior hybrids were available to farmers in commercial quantities.” The region where this breakthrough occurred was the Sixth Iowa Crop Reporting District [Griliches 1957b: 507 and Table II, p. 508]. The sixth district comprised Bremer, Black Hawk, Benton, Buchanan, Linn, Delaware, Jones, Dubuque, Jackson, and Clinton Counties, all in Iowa. 2 Paul David, in an insightful review, has criticized the Griliches approach “for lacking any real micro-level technology choice model” [David 2003:5]. Edwin Mansfield [1961] can be credited with supplying such a model to explain the logistic shape (although as David points out, Mansfield’s model is simply one of many formulations consistent with the data). Mansfield suggested that the probability that a non-user would switch to a new technology would be a function of the number of those in the immediate neighborhood who had already accepted the technology. This “contagion” model, borrowed from epidemiology, leads to the logistic diffusion curve. Bronwyn Hall [2004] provides a review of the theoretical literature on diffusion from both sociology and economics. 3 State level data on the percentage of corn acres planted to hybrids are available in various annual issues of Agricultural Statistics. I have relied on the volumes for 1945 (Table 46, p. 42), 1948 (Table 50, p. 48), 1950 (Table 49, p. 47), 1952 (Table 43, p. 40), 1954 (Table 38, p. 30), 1957 (Table 40, p. 39), 1959 (Table 43, p. 33), and 1961 (Table 43, p.33). 4 Griliches estimated the rate of return on hybrid corn research of at least 700 percent annually as of 1955 [Griliches 1958:419]. Richard Sutch, Adoption of Hybrid Corn Page 3 of 26 more famous, at least among plant scientists, than Griliches’s logistic, the rise in corn yields per acre is employed to suggest that hybrid corn was responsible for a biotechnological revolution that abruptly ended a sustained period of “biological stasis.” Figure 3 displays the “hockey stick” graph reproduced dozens of time in the scientific literature.5 The chart plots USDA statistics on corn yields per acre dating back to 1866. There was a remarkable stability in yields with no discernable trend before 1936.6 Thereafter yields began to increase and they have continued upward ever since. Yields per acre rose from an average of 25 bushels per acre before 1936 to 135 bushels per acre in the years 2000-2002 [Carter et al 2006: Series Da693-694], more than a five-fold increase. Perhaps too casually, this increase has been attributed (1) to the continuing adoption of the hybrid varieties between 1936 and 1960 and (2) to the continuing improvement of hybrid traits as new varieties were introduced between 1936 and 1989 [Duvick 1992]. I say “perhaps too casually,” because the introduction of hybrids was also accompanied by the increase use of synthetic nitrogen fertilizers, increased planting densities, and the adoption and improvements in planting and harvesting machinery. However, these developments were intimately interrelated. One of the hybrid traits introduced improved the plant’s ability to absorb nitrogen fertilizers and indeed the use of fertilizer was required to reach the potential of the hybrids. Similarly, the increased planting densities were possible only because of traits that reduced the plant’s requirements for full sunlight and that increased its resistance to lodging – the tendency of the plant to lie down or fall over when beaten down by the wind. Even then high density was possible only with the heavy application of fertilizer. Increased planting 5 As an indication of how ubiquitously Figure 3 appears, I note that a standard textbook on corn for plant scientists [Smith, Betrán, and Runge 2004] reproduces a version of this chart four times in four separate chapters [Troyer 2004: Chapter 1.4, Figure 32, p.